## File: README.md # InsightFace: 2D and 3D Face Analysis Project InsightFace project is mainly maintained by [Jia Guo](mailto:guojia@insightface.ai) and [Jiankang Deng](https://jiankangdeng.github.io/). For more information, please visit our website at [https://insightface.ai](https://insightface.ai) ## InsightFace 1.0 Update InsightFace 1.0 focuses on easier local evaluation and desktop usage: - The default Python package no longer builds the optional C++/Cython `face3d` extension, reducing local compiler requirements. - Added **InsightFace Evaluation Studio**, a cross-platform GUI Demo for Windows, macOS, and Linux. - The GUI supports local face recognition, enterprise model evaluation/reporting, and basic face swap trials. ## InsightFace Server The new [InsightFace Server](server/) provides a simple self-hosted Web UI, snake_case REST API, and Python client for detection, comparison, registration, and Person search. A single Linux x86_64 CPU or NVIDIA GPU container runs local ONNX Runtime inference and SQLite with operator-supplied, manifest-verified models. It is a simple AWS Rekognition alternative with accuracy-preserving INT8 embedding quantization and 50M+ image search on one RTX 5090 GPU. ## License The code of InsightFace is released under the MIT License. There is no limitation for both academic and commercial usage. The training data containing the annotation (and the models trained with these data) are available for non-commercial research purposes only. Both manual-downloading models from our github repo and auto-downloading models with our [python-library](python-package) follow the above license policy(which is for non-commercial research purposes only). `2025-11-24 Update:` 1. For inswapper series face swap models (e.g., inswapper_128.onnx/[inswapper-512-live](https://github.com/deepinsight/inswapper-512-live)), please contact [contact@insightface.ai](mailto:contact@insightface.ai) for licensing and additional support. 2. For open-sourced face recognition models (e.g., buffalo_l package), please contact [recognition-oss-pack@insightface.ai](mailto:recognition-oss-pack@insightface.ai) for licensing. 3. For advanced face recognition SDK and models (e.g., InspireFace SDK), please contact [contact@insightface.ai](mailto:contact@insightface.ai) for licensing and additional support. ## Top News **`2026-07-27`** [InsightFace Server](server/) Added a simple AWS Rekognition alternative with accuracy-preserving INT8 embedding quantization and 50M+ image search on one RTX 5090 GPU. **`2026-05-23`** `InsightFace 1.0` Added a cross-platform desktop GUI Demo for face recognition, enterprise evaluation, reports, and face swap trials, with a lighter default Python install that removes C++ build requirements. **`2025-11-18`** `[Picsi.ai]` Released Live Face Swap macOS & iOS App and updated [Picsi.ai](https://www.picsi.ai) services with our latest series of swap models (incl. [inswapper-512-live](https://github.com/deepinsight/inswapper-512-live)/Cyn/Dax). **`2024-05-04`** `[Picsi.ai]` Released [InspireFace](cpp-package/inspireface), a cross-platform C/C++ face recognition SDK. **`2022-08-12`**: We achieved Rank-1st of [Perspective Projection Based Monocular 3D Face Reconstruction Challenge](https://tianchi.aliyun.com/competition/entrance/531961/introduction) of [ECCV-2022 WCPA Workshop](https://sites.google.com/view/wcpa2022), [paper](https://arxiv.org/abs/2208.07142) and [code](reconstruction/jmlr). **`2021-10-29`**: We achieved 1st place on the [VISA track](https://pages.nist.gov/frvt/plots/11/visa.html) of [NIST-FRVT 1:1](https://pages.nist.gov/frvt/html/frvt11.html) by using Partial FC (Xiang An, Jiankang Deng, Jia Guo). ## ChangeLogs **`2026-07-27`** [InsightFace Server](server/) Added a simple AWS Rekognition alternative with accuracy-preserving INT8 embedding quantization and 50M+ image search on one RTX 5090 GPU. **`2025-11-18`** `[Picsi.ai]` Released Live Face Swap macOS & iOS App and updated [Picsi.ai](https://www.picsi.ai) services with our latest series of swap models (incl. [inswapper-live](https://github.com/deepinsight/inswapper-512-live)/Cyn/Dax). **`2024-05-04`** `[Picsi.ai]` Released [InspireFace](cpp-package/inspireface), a cross-platform C/C++ face recognition SDK. **`2024-04-17`**: [Monocular Identity-Conditioned Facial Reflectance Reconstruction](https://arxiv.org/abs/2404.00301) accepted by [CVPR-2024](https://cvpr.thecvf.com/Conferences/2024). **`2023-08-08`**: We released the implementation of [Generalizing Gaze Estimation with Weak-Supervision from Synthetic Views](https://arxiv.org/abs/2212.02997) at [reconstruction/gaze](reconstruction/gaze). **`2023-05-03`**: We have launched the ongoing version of wild face anti-spoofing challenge. See details [here](https://github.com/deepinsight/insightface/tree/master/challenges/cvpr23-fas-wild#updates). **`2023-02-13`**: We launch a large scale in the wild face anti-spoofing challenge on CVPR23 Workshop, see details at [challenges/cvpr23-fas-wild](challenges/cvpr23-fas-wild). **`2022-11-28`**: Single line code for facial identity swapping in our python packge ver 0.7, please check the example [here](examples/in_swapper). **`2022-10-28`**: [MFR-Ongoing](http://iccv21-mfr.com) website is refactored, please create issues if there's any bug. **`2022-09-22`**: Now we have [web-demos](web-demos): [face-localization](http://demo.insightface.ai:7007/), [face-recognition](http://demo.insightface.ai:7008/), and [face-swapping](http://demo.insightface.ai:7009/). **`2022-08-12`**: We achieved Rank-1st of [Perspective Projection Based Monocular 3D Face Reconstruction Challenge](https://tianchi.aliyun.com/competition/entrance/531961/introduction) of [ECCV-2022 WCPA Workshop](https://sites.google.com/view/wcpa2022), [paper](https://arxiv.org/abs/2208.07142) and [code](reconstruction/jmlr). **`2022-03-30`**: [Partial FC](https://arxiv.org/abs/2203.15565) accepted by CVPR-2022. **`2022-02-23`**: [SCRFD](detection/scrfd) accepted by [ICLR-2022](https://iclr.cc/Conferences/2022). **`2021-11-30`**: [MFR-Ongoing](challenges/mfr) challenge launched(same with IFRT), which is an extended version of [iccv21-mfr](challenges/iccv21-mfr). **`2021-10-29`**: We achieved 1st place on the [VISA track](https://pages.nist.gov/frvt/plots/11/visa.html) of [NIST-FRVT 1:1](https://pages.nist.gov/frvt/html/frvt11.html) by using Partial FC (Xiang An, Jiankang Deng, Jia Guo). **`2021-10-11`**: [Leaderboard](https://insightface.ai/mfr21) of [ICCV21 - Masked Face Recognition Challenge](challenges/iccv21-mfr) released. Video: [Youtube](https://www.youtube.com/watch?v=lL-7l5t6x2w), [Bilibili](https://www.bilibili.com/video/BV15b4y1h79N/). **`2021-06-05`**: We launch a [Masked Face Recognition Challenge & Workshop](challenges/iccv21-mfr) on ICCV 2021. ## Introduction [InsightFace](https://insightface.ai) is an open source 2D&3D deep face analysis toolbox, mainly based on PyTorch and MXNet. Please check our [website](https://insightface.ai) for detail. The master branch works with **PyTorch 1.6+** and/or **MXNet=1.6-1.8**, with **Python 3.x**. InsightFace efficiently implements a rich variety of state of the art algorithms of face recognition, face detection and face alignment, which optimized for both training and deployment. ## Quick Start Please start with our [python-package](python-package/), for testing detection, recognition and alignment models on input images. ### ArcFace Video Demo [](https://www.youtube.com/watch?v=y-D1tReryGA&t=81s) Please click the image to watch the Youtube video. For Bilibili users, click [here](https://www.bilibili.com/video/av38041494?from=search&seid=11501833604850032313). ## Projects The [page](https://insightface.ai/projects) on InsightFace website also describes all supported projects in InsightFace. You may also interested in some [challenges](https://insightface.ai/challenges) hold by InsightFace. ## Face Recognition ### Introduction In this module, we provide training data, network settings and loss designs for deep face recognition. The supported methods are as follows: - [x] [ArcFace_mxnet (CVPR'2019)](recognition/arcface_mxnet) - [x] [ArcFace_torch (CVPR'2019)](recognition/arcface_torch) - [x] [SubCenter ArcFace (ECCV'2020)](recognition/subcenter_arcface) - [x] [PartialFC_mxnet (CVPR'2022)](recognition/partial_fc) - [x] [PartialFC_torch (CVPR'2022)](recognition/arcface_torch) - [x] [VPL (CVPR'2021)](recognition/vpl) - [x] [Arcface_oneflow](recognition/arcface_oneflow) - [x] [ArcFace_Paddle (CVPR'2019)](recognition/arcface_paddle) Commonly used network backbones are included in most of the methods, such as IResNet, MobilefaceNet, MobileNet, InceptionResNet_v2, DenseNet, etc.. ### Datasets The training data includes, but not limited to the cleaned MS1M, VGG2 and CASIA-Webface datasets, which were already packed in MXNet binary format. Please [dataset](recognition/_datasets_) page for detail. ### Evaluation We provide standard IJB and Megaface evaluation pipelines in [evaluation](recognition/_evaluation_) ### Pretrained Models **Please check [Model-Zoo](https://github.com/deepinsight/insightface/wiki/Model-Zoo) for more pretrained models.** ### Third-party Re-implementation of ArcFace - TensorFlow: [InsightFace_TF](https://github.com/auroua/InsightFace_TF) - TensorFlow: [tf-insightface](https://github.com/AIInAi/tf-insightface) - TensorFlow:[insightface](https://github.com/Fei-Wang/insightface) - PyTorch: [InsightFace_Pytorch](https://github.com/TreB1eN/InsightFace_Pytorch) - PyTorch: [arcface-pytorch](https://github.com/ronghuaiyang/arcface-pytorch) - Caffe: [arcface-caffe](https://github.com/xialuxi/arcface-caffe) - Caffe: [CombinedMargin-caffe](https://github.com/gehaocool/CombinedMargin-caffe) - Tensorflow: [InsightFace-tensorflow](https://github.com/luckycallor/InsightFace-tensorflow) - TensorRT: [wang-xinyu/tensorrtx](https://github.com/wang-xinyu/tensorrtx) - TensorRT: [InsightFace-REST](https://github.com/SthPhoenix/InsightFace-REST) - ONNXRuntime C++: [ArcFace-ONNXRuntime](https://github.com/DefTruth/lite.ai.toolkit/blob/main/lite/ort/cv/glint_arcface.cpp) - ONNXRuntime Go: [arcface-go](https://github.com/jack139/arcface-go) - MNN: [ArcFace-MNN](https://github.com/DefTruth/lite.ai.toolkit/blob/main/lite/mnn/cv/mnn_glint_arcface.cpp) - TNN: [ArcFace-TNN](https://github.com/DefTruth/lite.ai.toolkit/blob/main/lite/tnn/cv/tnn_glint_arcface.cpp) - NCNN: [ArcFace-NCNN](https://github.com/DefTruth/lite.ai.toolkit/blob/main/lite/ncnn/cv/ncnn_glint_arcface.cpp) ## Face Detection ### Introduction In this module, we provide training data with annotation, network settings and loss designs for face detection training, evaluation and inference. The supported methods are as follows: - [x] [RetinaFace (CVPR'2020)](detection/retinaface) - [x] [SCRFD (Arxiv'2021)](detection/scrfd) - [x] [blazeface_paddle](detection/blazeface_paddle) [RetinaFace](detection/retinaface) is a practical single-stage face detector which is accepted by [CVPR 2020](https://openaccess.thecvf.com/content_CVPR_2020/html/Deng_RetinaFace_Single-Shot_Multi-Level_Face_Localisation_in_the_Wild_CVPR_2020_paper.html). We provide training code, training dataset, pretrained models and evaluation scripts. [SCRFD](detection/scrfd) is an efficient high accuracy face detection approach which is initialy described in [Arxiv](https://arxiv.org/abs/2105.04714). We provide an easy-to-use pipeline to train high efficiency face detectors with NAS supporting. ## Face Alignment ### Introduction In this module, we provide datasets and training/inference pipelines for face alignment. Supported methods: - [x] [SDUNets (BMVC'2018)](alignment/heatmap) - [x] [SimpleRegression](alignment/coordinate_reg) [SDUNets](alignment/heatmap) is a heatmap based method which accepted on [BMVC](http://bmvc2018.org/contents/papers/0051.pdf). [SimpleRegression](alignment/coordinate_reg) provides very lightweight facial landmark models with fast coordinate regression. The input of these models is loose cropped face image while the output is the direct landmark coordinates. ## Citation If you find *InsightFace* useful in your research, please consider to cite the following related papers: ``` /* Detailed source-code truncated for AI context efficiency. */ ``` ## Contributing Main contributors: - [Jia Guo](https://github.com/nttstar), ``guojia[at]gmail.com`` - [Jiankang Deng](https://github.com/jiankangdeng) ``jiankangdeng[at]gmail.com`` - [Xiang An](https://github.com/anxiangsir) ``anxiangsir[at]gmail.com`` - [Jack Yu](https://github.com/szad670401) ``jackyu961127[at]gmail.com`` - [Baris Gecer](https://barisgecer.github.io/) ``barisgecer[at]msn.com`` `` --- ## File: alignment/heatmap/README.md We provide our implementation of ``Stacked Dense U-Nets with Dual Transformers for Robust Face Alignment`` here at [BMVC](http://bmvc2018.org/contents/papers/0051.pdf) or link at [Arxiv](https://arxiv.org/abs/1812.01936). We also provide some popular heatmap based approaches like stacked hourglass, etc.. You can define different loss-type/network structure/dataset in ``config.py``(from ``sample_config.py``). For example, by default, you can train our approach by ``train.py --network sdu`` or train hourglass network by ``train.py --network hourglass``. 2D training/validation dataset is now available at [baiducloud](https://pan.baidu.com/s/1kdquiIGTlK7l26SPWO_cmw) or [dropbox](https://www.dropbox.com/s/por6mbguegmywo6/bmvc_sdu_data2d.zip?dl=0) 3D training/validation dataset is now available at [baiducloud](https://pan.baidu.com/s/1VjFWm6eEtIqGKk92GE2rgw) or [dropbox](https://www.dropbox.com/s/tjze176lh76nciw/bmvc_sdu_data3d.zip?dl=0) --- ## File: alignment/coordinate_reg/README.md ### Introduction Here we provide some lightweight facial landmark models with fast coordinate regression. The input of these models is loose cropped face image while the output is the direct landmark coordinates. ### Pretrained Models - **Model ``2d106det``** **2021.07: We now support model inference by our `insightface` python package, please check [image_infer.py](image_infer.py) for detail.** Given face detection bounding box, predict 2d-106 landmarks. Mainly used for static image inference. Backbone: MobileNet-0.5, size 5MB. Input: size 192x192, loose cropped detection bounding-box. Download link: [baidu cloud](https://pan.baidu.com/s/10m5GmtNV5snynDrq3KqIdg) (code: ``lqvv``) [google drive](https://drive.google.com/file/d/13Pz8mH-a1s7RXpq_jFUXxaqCpDUE0oSr/view?usp=sharing) - **Model ``2d106track``** Given landmarks bounding box, predict 2d-106 landmarks. Used for video landmarks tracking. Download link: coming soon ### Visualization ### FAQ --- ## File: alignment/_datasets_/README.md # Face Alignment Datasets (Updating) ## Training Datasets ### Menpo2D-Train https://ibug.doc.ic.ac.uk/resources/2nd-facial-landmark-tracking-competition-menpo-ben/ ### 300W-Train https://ibug.doc.ic.ac.uk/resources/300-W/ ### LFPW https://neerajkumar.org/databases/lfpw/ ### Helen http://www.ifp.illinois.edu/~vuongle2/helen/ ### AFW ### AFLW https://www.tugraz.at/institute/icg/research/team-bischof/learning-recognition-surveillance/downloads/aflw ### FDDB ### Face Synthetics https://github.com/microsoft/FaceSynthetics ### 300W-LP (3D annotation) http://www.cbsr.ia.ac.cn/users/xiangyuzhu/projects/3DDFA/main.htm ## Test Datasets ### 300W-Test https://ibug.doc.ic.ac.uk/resources/300-W/ ### COFW http://www.vision.caltech.edu/xpburgos/ICCV13/#dataset ### Menpo2D-Test https://ibug.doc.ic.ac.uk/resources/2nd-facial-landmark-tracking-competition-menpo-ben/ ### AFLW2000-3D (3D annotation) http://www.cbsr.ia.ac.cn/users/xiangyuzhu/projects/3DDFA/main.htm --- ## File: alignment/synthetics/README.md # Introduction We provide training and testing tools on synthetics data. ## Dataset ### Training dataset Download `Face Synthetics dataset` from [https://github.com/microsoft/FaceSynthetics](https://github.com/microsoft/FaceSynthetics) and put it somewhere. Then use [tools/prepare_synthetics.py](tools/prepare_synthetics.py) for training data preparation. ### Testing dataset [300-W](https://ibug.doc.ic.ac.uk/resources/300-W/) ## Pretrained Model [ResNet50d](https://drive.google.com/file/d/1kNP7qEl3AYNbaHFUg_ZiyRB1CtfDWXR4/view?usp=sharing) ## Train and Test ### Prerequisites - pytorch_lightning - timm - albumentations ### Training `` python -u trainer_synthetics.py `` which uses `resnet50d` as backbone by default, please check the [code](trainer_synthetics.py) for detail. ### Testing Please check [test_synthetics.py](test_synthetics.py) for detail. ## Result Visualization(3D 68 Keypoints) --- ## File: alignment/README.md ## Face Alignment ## Introduction These are the face alignment methods of [InsightFace](https://insightface.ai) ### Datasets Please refer to [datasets](_datasets_) page for the details of face alignment datasets used for training and evaluation. ### Evaluation Please refer to [evaluation](_evaluation_) page for the details of face alignment evaluation. ## Methods Supported methods: - [x] [SDUNets (BMVC'2018)](heatmap) - [x] [SimpleRegression](coordinate_reg) - [x] [Alignment By Face Synthetics](synthetics) ## Contributing We appreciate all contributions to improve the face alignment model zoo of InsightFace. --- ## File: attribute/_datasets_/README.md # Face Attribute Datasets (Updating) ## Training Datasets ### CelebA https://mmlab.ie.cuhk.edu.hk/projects/CelebA.html ## Test Datasets --- ## File: attribute/README.md ## Face Attribute ## Introduction These are the face attribute methods of [InsightFace](https://insightface.ai) ## Methods Supported methods: - [x] [Gender_Age](gender_age) ## Contributing We appreciate all contributions to improve the face attribute module of InsightFace. --- ## File: challenges/mfr/README.md # MFR Ongoing 💡 **Due to GPU resource constraints, we are temporarily taking this service offline. If any companies or individuals are interested in hosting this service, please contact me via email: ``guojia AT insightface.ai``** This is the ongoing version of [ICCV-2021 Masked Face Recognition Challenge & Workshop(MFR)](https://ibug.doc.ic.ac.uk/resources/masked-face-recognition-challenge-workshop-iccv-21/). We also extend it to involve some public available and popular benchmarks such as IJBC, LFW, CFPFP and AgeDB. (:bulb: :bulb: Once you find the name **IFRT** which is *InsightFace Recognition Test* in short anywhere, it is the same as MFR-Ongoing.) For detail, please check our ICCV 2021 workshop [paper](https://openaccess.thecvf.com/content/ICCV2021W/MFR/papers/Deng_Masked_Face_Recognition_Challenge_The_InsightFace_Track_Report_ICCVW_2021_paper.pdf). More information about the workshop challenge can be found [here](../iccv21-mfr), for reference. **MFR** testset consists of non-celebrities so we can ensure that it has very few overlap with public available face recognition training set, such as MS1M and CASIA as they mostly collected from online celebrities. As the result, we can evaluate the FAIR performance for different algorithms. In recent changes, we also add public available popular benchmarks such as IJBC, LFW, CFPFP, AgeDB into **MFR-Ongoing**. Current submission server link: [http://iccv21-mfr.com/](http://iccv21-mfr.com/) For any question, please send email to `insightface.challenge AT gmail.com` ## Testsets In MFR-Ongoing, we will evaluate the accuracy of following testsets: * **Accuracy between masked and non-masked faces.** * **Accuracy among children(2~16 years old).** * **Accuracy of globalised multi-racial benchmarks.** We ensure that there's no overlap between the above testsets and public available training datasets, as they are not collected from online celebrities. We also evaluate below public available popular benchmarks: * **IJBC under FAR<=e-5 and FAR<=e-4.** * **Some 1:1 verification testsets, such as LFW, CFPFP, AgeDB-30.** ### ``Mask test-set:`` Mask testset contains 6,964 identities, 6,964 masked images and 13,928 non-masked images. There are totally 13,928 positive pairs and 96,983,824 negative pairs. Click to check the sample images(here we manually blur it to protect privacy) ### ``Children test-set:`` Children testset contains 14,344 identities and 157,280 images. There are totally 1,773,428 positive pairs and 24,735,067,692 negative pairs. Click to check the sample images(here we manually blur it to protect privacy) ### ``Multi-racial test-set (MR in short):`` The globalised multi-racial testset contains 242,143 identities and 1,624,305 images. | Race-Set | Identities | Images | Positive Pairs | Negative Pairs | | ------- | ---------- | ----------- | ----------- | ----------- | | African | 43,874 | 298,010 | 870,091 | 88,808,791,999 | | Caucasian | 103,293 | 697,245 | 2,024,609 | 486,147,868,171 | | Indian | 35,086 | 237,080 | 688,259 | 56,206,001,061 | | Asian | 59,890 | 391,970 | 1,106,078 | 153,638,982,852 | | **ALL** | **242,143** | **1,624,305** | **4,689,037** | **2,638,360,419,683** | Click to check the sample images(here we manually blur it to protect privacy) ## Evaluation Metric For ``Mask`` set, TAR is measured on mask-to-nonmask 1:1 protocal, with FAR less than 0.0001(e-4). For ``Children`` set, TAR is measured on all-to-all 1:1 protocal, with FAR less than 0.0001(e-4). For multi-racial sets, TAR is measured on all-to-all 1:1 protocal, with FAR less than 0.000001(e-6). For IJBC and verification test-set, we use the most common test protocal. Participants are ordered in terms of highest scores across two datasets: **TAR@Mask** and **TAR@MR-All**, by the formula of ``0.25 * TAR@Mask + 0.75 * TAR@MR-All``. ## Baselines **``2021.04.25``** We made a clean on East Asian subset, by removing children images. **``2021.04.27``** Add onnx download links. | Backbone | Dataset | Method | Mask | Children | African | Caucasian | South Asian | East Asian | All | size(mb) | infer(ms) | link | |------------|------------|------------|--------|----------|---------|-----------|-------------|------------|--------|----------|-----------|-----------| | R100 | Casia | ArcFace | 26.623 | 30.359 | 39.666 | 53.933 | 47.807 | 21.572 | 42.735 | 248.904 | 7.073 | [download](https://1drv.ms/u/s!AswpsDO2toNKrUJpk8zC61HVN7Kg?e=zE9JDd) | | R100 | MS1MV2 | ArcFace | 65.767 | 60.496 | 79.117 | 87.176 | 85.501 | 55.807 | 80.725 | 248.904 | 7.028 | [download](https://1drv.ms/u/s!AswpsDO2toNKrUTlYEHJCHg3UYM-?e=ihxMpS) | | R18 | MS1MV3 | ArcFace | 47.853 | 41.047 | 62.613 | 75.125 | 70.213 | 43.859 | 68.326 | 91.658 | 1.856 | [download](https://1drv.ms/u/s!AswpsDO2toNKrTxlT6w1Jo02yzSh?e=KDhFAA) | | R34 | MS1MV3 | ArcFace | 58.723 | 55.834 | 71.644 | 83.291 | 80.084 | 53.712 | 77.365 | 130.245 | 3.054 | [download](https://1drv.ms/u/s!AswpsDO2toNKrT2O5pgyVtwnjeMq?e=16S8LI) | | R50 | MS1MV3 | ArcFace | 63.850 | 60.457 | 75.488 | 86.115 | 84.305 | 57.352 | 80.533 | 166.305 | 4.262 | [download](https://1drv.ms/u/s!AswpsDO2toNKrUUWd5i3a5OlFpM_?e=ExBDBN) | | R100 | MS1MV3 | ArcFace | 69.091 | 66.864 | 81.083 | 89.040 | 88.082 | 62.193 | 84.312 | 248.590 | 7.031 | [download](https://1drv.ms/u/s!AswpsDO2toNKrUPwyqWvNXUlNd3P?e=pTLw9A) | | R18 | Glint360K | ArcFace | 53.317 | 48.113 | 68.230 | 80.575 | 75.852 | 47.831 | 72.074 | 91.658 | 2.013 | [download](https://1drv.ms/u/s!AswpsDO2toNKrT5ey4lCqFzlpzDd?e=VWP28J) | | R34 | Glint360K | ArcFace | 65.106 | 65.454 | 79.907 | 88.620 | 86.815 | 60.604 | 83.015 | 130.245 | 3.044 | [download](https://1drv.ms/u/s!AswpsDO2toNKrUBcgGkiuUS11Hsd?e=ISGDnP) | | R50 | Glint360K | ArcFace | 70.233 | 69.952 | 85.272 | 91.617 | 90.541 | 66.813 | 87.077 | 166.305 | 4.340 | [download](https://1drv.ms/u/s!AswpsDO2toNKrT8jbvHxjqCY0d08?e=igfdrd) | | R100 | Glint360K | ArcFace | 75.567 | 75.202 | 89.488 | 94.285 | 93.434 | 72.528 | 90.659 | 248.590 | 7.038 | [download](https://1drv.ms/u/s!AswpsDO2toNKrUFgLEIj-mnkb51b?e=vWqy2q) | | - | *Private* | insightface-000 of frvt | 97.760 | 93.358 | 98.850 | 99.372 | 99.058 | 87.694 | 97.481 | - | - | - | (MS1M-V2 means MS1M-ArcFace, MS1M-V3 means MS1M-RetinaFace). Inference time in above table was evaluated on Tesla V100 GPU, using onnxruntime-gpu==1.6. ## Rules 1. We have two tracks, academic and unconstrained. 2. Please **DO NOT** register the account with messy or random characters(for both username and organization). 3. **For academic submissions, we recommend to set the username as the name of your proposed paper or method. Orgnization hiding is not allowed(or the score will be banned) for this track but you can set the submission as private. You can also create multiple accounts, one account for one method.** 4. Right now we only support 112x112 input, so make sure that the submission model accepts the correct input shape(['*',3,112,112]), in RGB order. Add an interpolate operator into the first layer of the submission model if you need a different input resolution. 5. Participants submit onnx model, then get scores by our online evaluation. 6. Matching score is measured by cosine similarity. 7. **Online evaluation server uses onnxruntime-gpu==1.8, cuda==11.1, cudnn==8.0.5, GPU is RTX3090.** 8. Any float-16 model weights is prohibited, as it will lead to incorrect model size estimiation. 9. Please use ``onnx_helper.py`` to check whether the model is valid. 10. Leaderboard is now ordered in terms of highest scores across two datasets: **TAR@Mask** and **TAR@MR-All**, by the formula of ``0.25 * TAR@Mask + 0.75 * TAR@MR-All``. ## Submission Guide 1. Participants must package the onnx model for submission using ``zip xxx.zip model.onnx``. 2. Each participant can submit three times a day at most. 3. Please sign-up with the real organization name. You can hide the organization name in our system if you like(not allowed for academic track). 4. You can decide which submission to be displayed on the leaderboard by clicking 'Set Public' button. 5. Please click 'sign-in' on submission server if find you're not logged in. --- ## File: challenges/iccv21-mfr/README.md # InsightFace Track of ICCV21-MFR ## NEWS **``2021-11-30``** [MFR-Ongoing](../mfr) is now available. **``2021-10-26``** Please send the onnx models to us(insightface.challenge[at]gmail.com) if you want to test the MFR accuracy before our system rebooting(may be in Nov.). **``2021-10-11``** [Final Leaderboard](https://insightface.ai/mfr21) **``2021-10-04``** Please fix the public leaderboard scores before ``2021-10-05 20:00(UTC+8 Time)`` **``2021-07-16``** Implicit batch inference is prohibited. For example, inserting some data-related OPs to onnx graph to enable automatic flip-test is not allowed(or similar ideas). We will check it after submission closed, to ensure fairness. **``2021-06-17``** Participants are now ordered in terms of highest scores across two datasets: **TAR@Mask** and **TAR@MR-All**, by the formula of ``0.25 * TAR@Mask + 0.75 * TAR@MR-All``. ## Introduction The Masked Face Recognition Challenge & Workshop(MFR) will be held in conjunction with the International Conference on Computer Vision (ICCV) 2021. [Workshop-Homepage](https://ibug.doc.ic.ac.uk/resources/masked-face-recognition-challenge-workshop-iccv-21/). There're InsightFace track here and [Webface260M](https://www.face-benchmark.org/challenge.html) track(with larger training set) in this workshop. [**Challenge Leaderboard**](https://insightface.ai/mfr21) Submission server link: [http://iccv21-mfr.com/](http://iccv21-mfr.com/) An alternative submission server for Non-Chinese users: [http://124.156.136.55/](http://124.156.136.55/) ### Discussion group: WeChat: QQ Group: 711302608, *answer: github* Online issue discussion: [https://github.com/deepinsight/insightface/issues/1564](https://github.com/deepinsight/insightface/issues/1564) ## Testsets for insightface track In this challenge, we will evaluate the accuracy of following testsets: * Accuracy between masked and non-masked faces. * Accuracy among children(2~16 years old). * Accuracy of globalised multi-racial benchmarks. We ensure that there's no overlap between these testsets and public available training datasets, as they are not collected from online celebrities. Our test datasets mainly comes from [IFRT](../ifrt). ### ``Mask test-set:`` Mask testset contains 6,964 identities, 6,964 masked images and 13,928 non-masked images. There are totally 13,928 positive pairs and 96,983,824 negative pairs. Click to check the sample images(here we manually blur it to protect privacy) ### ``Children test-set:`` Children testset contains 14,344 identities and 157,280 images. There are totally 1,773,428 positive pairs and 24,735,067,692 negative pairs. Click to check the sample images(here we manually blur it to protect privacy) ### ``Multi-racial test-set (MR in short):`` The globalised multi-racial testset contains 242,143 identities and 1,624,305 images. | Race-Set | Identities | Images | Positive Pairs | Negative Pairs | | ------- | ---------- | ----------- | ----------- | ----------- | | African | 43,874 | 298,010 | 870,091 | 88,808,791,999 | | Caucasian | 103,293 | 697,245 | 2,024,609 | 486,147,868,171 | | Indian | 35,086 | 237,080 | 688,259 | 56,206,001,061 | | Asian | 59,890 | 391,970 | 1,106,078 | 153,638,982,852 | | **ALL** | **242,143** | **1,624,305** | **4,689,037** | **2,638,360,419,683** | Click to check the sample images(here we manually blur it to protect privacy) ## Evaluation Metric For ``Mask`` set, TAR is measured on mask-to-nonmask 1:1 protocal, with FAR less than 0.0001(e-4). For ``Children`` set, TAR is measured on all-to-all 1:1 protocal, with FAR less than 0.0001(e-4). For other sets, TAR is measured on all-to-all 1:1 protocal, with FAR less than 0.000001(e-6). Participants are ordered in terms of highest scores across two datasets: **TAR@Mask** and **TAR@MR-All**, by the formula of ``0.25 * TAR@Mask + 0.75 * TAR@MR-All``. ## Baselines | Backbone | Dataset | Method | Mask | Children | African | Caucasian | South Asian | East Asian | MR-All | size(mb) | infer(ms) | link | |------------|------------|------------|--------|----------|---------|-----------|-------------|------------|--------|----------|-----------|-----------| | R100 | Casia | ArcFace | 26.623 | 30.359 | 39.666 | 53.933 | 47.807 | 21.572 | 42.735 | 248.904 | 7.073 | [download](https://1drv.ms/u/s!AswpsDO2toNKrUJpk8zC61HVN7Kg?e=zE9JDd) | | R100 | MS1MV2 | ArcFace | 65.767 | 60.496 | 79.117 | 87.176 | 85.501 | 55.807 | 80.725 | 248.904 | 7.028 | [download](https://1drv.ms/u/s!AswpsDO2toNKrUTlYEHJCHg3UYM-?e=ihxMpS) | | R18 | MS1MV3 | ArcFace | 47.853 | 41.047 | 62.613 | 75.125 | 70.213 | 43.859 | 68.326 | 91.658 | 1.856 | [download](https://1drv.ms/u/s!AswpsDO2toNKrTxlT6w1Jo02yzSh?e=KDhFAA) | | R34 | MS1MV3 | ArcFace | 58.723 | 55.834 | 71.644 | 83.291 | 80.084 | 53.712 | 77.365 | 130.245 | 3.054 | [download](https://1drv.ms/u/s!AswpsDO2toNKrT2O5pgyVtwnjeMq?e=16S8LI) | | R50 | MS1MV3 | ArcFace | 63.850 | 60.457 | 75.488 | 86.115 | 84.305 | 57.352 | 80.533 | 166.305 | 4.262 | [download](https://1drv.ms/u/s!AswpsDO2toNKrUUWd5i3a5OlFpM_?e=ExBDBN) | | R100 | MS1MV3 | ArcFace | 69.091 | 66.864 | 81.083 | 89.040 | 88.082 | 62.193 | 84.312 | 248.590 | 7.031 | [download](https://1drv.ms/u/s!AswpsDO2toNKrUPwyqWvNXUlNd3P?e=pTLw9A) | | R18 | Glint360K | ArcFace | 53.317 | 48.113 | 68.230 | 80.575 | 75.852 | 47.831 | 72.074 | 91.658 | 2.013 | [download](https://1drv.ms/u/s!AswpsDO2toNKrT5ey4lCqFzlpzDd?e=VWP28J) | | R34 | Glint360K | ArcFace | 65.106 | 65.454 | 79.907 | 88.620 | 86.815 | 60.604 | 83.015 | 130.245 | 3.044 | [download](https://1drv.ms/u/s!AswpsDO2toNKrUBcgGkiuUS11Hsd?e=ISGDnP) | | R50 | Glint360K | ArcFace | 70.233 | 69.952 | 85.272 | 91.617 | 90.541 | 66.813 | 87.077 | 166.305 | 4.340 | [download](https://1drv.ms/u/s!AswpsDO2toNKrT8jbvHxjqCY0d08?e=igfdrd) | | R100 | Glint360K | ArcFace | 75.567 | 75.202 | 89.488 | 94.285 | 93.434 | 72.528 | 90.659 | 248.590 | 7.038 | [download](https://1drv.ms/u/s!AswpsDO2toNKrUFgLEIj-mnkb51b?e=vWqy2q) | | - | *Private* | insightface-000 of frvt | 97.760 | 93.358 | 98.850 | 99.372 | 99.058 | 87.694 | 97.481 | - | - | - | (MS1M-V2 means MS1M-ArcFace, MS1M-V3 means MS1M-RetinaFace). Inference time was evaluated on Tesla V100 GPU, using onnxruntime-gpu==1.6. ## Rules 1. We have two sub-tracks, determined by the size of training dataset and inference time limitation. * Sub-Track A: Use MS1M-V3 as training set, download: [ref-link](https://github.com/deepinsight/insightface/blob/master/recognition/_datasets_/README.md#ms1m-retinaface), feature length must <= 512, and the inference time must <= 10ms on Tesla V100 GPU. * Sub-Track B: Use Glint360K as training set, download: [ref-link](https://github.com/deepinsight/insightface/blob/master/recognition/_datasets_/README.md#deepglint-181k-ids675m-images-8), feature length must <= 1024, and the inference time must <= 20ms on Tesla V100 GPU. 2. Training set and testing set are both aligned to 112x112, re-alignment is prohibited. 3. Mask data-augmentation is allowed, such as [this](../../recognition/_tools_). The applied mask augmentation tool should be reproducible. 4. External dataset and pretrained models are both prohibited. 5. Participants submit onnx model, then get scores by our online evaluation. Test images are invisible. 6. Matching score is measured by cosine similarity. 7. Model size must <= 1GB. 8. The input shape of submission model should equal to 3x112x112 (RGB order). 9. Online evaluation server uses onnxruntime-gpu==1.6, cuda==10.2, cudnn==8.0.5. 10. Any float-16 model weights is prohibited, as it will lead to incorrect model size estimiation. 11. Please use ``onnx_helper.py`` to check whether the model is valid. 12. Participants are now ordered in terms of highest scores across two datasets: **TAR@Mask** and **TAR@MR-All**, by the formula of ``0.25 * TAR@Mask + 0.75 * TAR@MR-All``. 13. Top-ranked participants should provide their solutions and codes to ensure their validity after submission closed. ## Tutorial 1. ArcFace-PyTorch (with Partial-FC), [code](../../recognition/arcface_torch), [tutorial-cn](tutorial_pytorch_cn.md) 2. OneFlow, [code](../../recognition/oneflow_face) 3. MXNet, [code](../../recognition/arcface_mxnet) ## Submission Guide 1. Participants must package the onnx model for submission using ``zip xxx.zip model.onnx``. 2. Each participant can submit three times a day at most. 3. Please sign-up with the real organization name. You can hide the organization name in our system if you like. 4. You can decide which submission to be displayed on the leaderboard by clicking 'Set Public' button. 5. Please click 'sign-in' on submission server if find you're not logged in. Server link: [http://iccv21-mfr.com/](http://iccv21-mfr.com/) ## Timelines * 1 June - Release of the training data, baseline solutions and testing leader-board * 1 October - Stop leader-board submission (11:59 PM Pacific Time) * 7 October - Winners notification ## Sponsors (in alphabetical order) **[DeepGlint](http://www.deepglint.com/)** **[Kiwi Tech](http://www.kiwiar.com)** **[OneFlow](https://www.oneflow.org)** **[Zoloz](https://www.zoloz.com)** ## Bonus Share | | Sub-Track A | Sub-Track B | | --------- | ------- | ------- | | 1st place | 30% | 30% | | 2nd place | 15% | 15% | | 3rd place | 5% | 5% |